Page:Miscellaneousbot01brow.djvu/580

 562 ON THE RELATIVE POSITION OF

derived from the veiy remarkable dehiscence of the capsule into six valves. But I have elsewhere pointed out cases where an analogous dehiscence occurs, in which, however, a similar composition has never been supposed to exist : and if the presence of six vascular cords in sections of the ova- rium be likewise adduced in favour of the opinion, I may add that I have in the same place remarked that these vascular bundles belong not to the ovarium only, but also to the perianthium and stamina, and are equally observable in other families w^ith adherent ovarium, as Iridece, in which a similar composition has never been inferred.

With regard to the second family, in which Mr. Lindley believes the disk of the carpel to be ovuliferous, namely, Orohancliece, I find no other argument advanced in support of this view than that derived from the bursting of the capsule into two lateral valves ; but an opinion founded on dehiscence only may be said to be a mere begging of the question ; division through the axis of carpels, especially in the families related to OrobanchecB, being nearly as common as separation of their margins. In this family also, as in OrcJndecB, the placentae are double, an argument in favour of their submarginal origin : and although, whether the carpels be regarded as lateral, or anterior and posterior, the placentae are not strictly marginal, yet there are other fami- lies where a similar position of placentae is found, but in which the structure assumed in this hypothesis has never been suspected. As to the supposed affinity of Orohanchem with GentianecB, which might be adduced in support of this view, as far as it is founded on the assumed agreement of the two orders in the lateral position of their carpels, the argument, even if correct, would hardly be conclusive ; for in GentianecB there is at least one genus having quadrifid and another with quinquefid flowers, in which the carpels are 112] not lateral, but anterior and posterior, as I beheve them to be in Orohanchece ; nor has it ever been supposed that in GentianecB the disk or axis is ovuliferous.

In the account now given of the modifications of ovarium and stigma, I have, in conformity with the ordinary language of botanists, employed the term confluence, by which, how-

�� �